Techniques taken from neural network evolution, which allow for efficient optimization of thousands of parameters simultaneously will be applied to the optimization of aerodynamic designs. Divergent search techniques, which ensure a diversity of designs, will play a key role. In optimization, these approaches will be used to encourage a fuller exploration of possible solutions, prevent premature convergence, and illuminate relationships between features and performance. A wider array of high-performing solutions is expected to aid in creation of the data-driven approximative aerodynamics models necessary to cope with the high computational cost of fluid dynamics simulations. Research will focus on design in the context of aerodynamics, though the resulting system could be applied to a variety of automated engineering problems.